BlueXP is now NetApp Console
Monitor and run hybrid cloud data services
So I'm Eric Khan, VP of product management at NetApp. Thank you for taking the time to join us. This is 1335 all about why Google Cloud is a more powerful, innovative cloud storage foreveryone. On stage with me, I'll have a chance to go through some of our background, but I also have Sean and Guna, which I'll introduce in a second. So you've all seen the confidentiality, right? So, uh, a couple things I want to say about the partnership. It's one that started quite some time ago and I think the innovation that we've done there's a lot that you could see now but it's important to realize that behind the scenes under the covers uh the teams have been working very hard for many years>> and it starts in 2019 we launched a cloud volume service in Google it started as a marketplace offering and you could see there's people here that have been driving that from a business point of view and it's been growing and scaling right so from that perspective we've been innovating on a couple fronts. One is on features and features include things like NFS, CMEC, uh, Kerros and then on top of that it's also been in deployment and coverage. We are now in 26 regions. We started off in a handful and you'll see that pattern repeating itself over and over again, right? And then the third dimension has really been on workload expansion. What started off as fileshares quickly it grew to include things like VMware VDI as well as customers are running SAP fileshares and SA starting to run SAP HANA. So all of that is to say it started quite some time ago but there's been a lot of exciting things right we've won partner of the year several times and then this most rec recent year with Google we were partner of the year in two categories the first is in storage infrastructure and the second one was as a marketplace offering right and then on stage with me today I have the leaders who have been driving the most recent launch which is Google cloud NetApp volumes as you know released as GA in October and it's I'm sorry not October August and it's quickly expanding into regions and we'll go through some of the use cases but in from a NetApp perspective it's a firstparty Google product that enables us to work very closely and co-develop going forward right in terms of the workloads there's a lot that you'll see in common with some of the other use cases but it starts with Windows right in Google customers are lift and shifting they're moving their data center workloads to the cloud NetApp supports Windows and as you know we support things like dual pro protocol so a single volume can support SMB and NFS on top of that there's file share so Linux file shares home directories and so on and then we mentioned SAP and then VDI is important as well as VMware the importance is not just the workloads but it's also reducing the toil so all of this is exposed as a volume service customers just consume they just consume a volume they get a mount and from then on they don't have to worry about managing some set of VMs that are hosting a fileshare. they get to scale independently the compute and the storage and all that management is taken care for them right so it's the beauty of being a volume service and then from a innovation point of view uh I'll give a high level perspective and then I'll ask some of my uh panelists to go into deeper detail it's that there is both an innovation in terms of the integrated experience so Sean will talk a lot about it but in our integrations now you can get the volume service or GCNV Google cloud NetF volumes straight from the command line. So G-Cloud tools, the UI billing is integrated. So it's all integrated as you would expect a Googleproduct and on top of that we're expanding the workloads. So I mentioned VMware uh certification which I'll defer to the others to talk more about and VDI. So those are all expanding and then this is more directional so you can see where we're going is we're also adding things like container support that's in the road map. And then one thing that customers have been bringing up is how to integrate with BigQuery. So we've been toying with certain ideas about how we can expose BigQuery and make that available to all the customers. So with that, let me start introducing my two panelists. Right. The first is Sean Darington. So Sean is GPM of Google Cloud Storage. Uh Sean, just to give him give some background because I've known Sean for some time. Uh he started off as a storage analyst. Maybe I should stand in front. He started off as a storage analyst at Metag Group which later was acquired by Gartner. From there he went on to do cloud product management and storage product management at semantic and then uh a startup. All of us have that in common. We've all done some storage startups. So we're all itching for pain. And he then went on to join Google running their uh cloud storage. So maybe I'll start with one question for Sean and then as I introduce uh Sean maybe you could share with the group two things. one is uh why has Google chosen to partner with NetApp for GCNV and also what does GCNV represent in your portfolio?Yeah. So, um, thanks. It's great to be here. And, uh, you know, one of the things that we've been looking atGoogle is we have an open ecosystem, right? So, we develop a lot of products ourselves and deliver them, um, as firstparty services, but we also have been working with partners uh, over the years and Eric mentioned the last uh, three and four with NetApp cloud volume service as well as uh, NetApp on. But what we realized was that sometimes transactions through the marketplace through partners is actually a barrier to customers adopting a product. uh they oftentimes want a product that is from the provider themselves. Uh and that provides the seamless experience not only from just the purchasing and procurement but also support and everything else that we've uh done to integrate this into a firstparty offering. And as we looked at the portfolio, you know, NetApp has a track record over the last couple of decades of being a leader in file services for on premises as well as I would also say within hypers scales not only within Google cloud but also AWS and Amazon. Uh and so we really looked at that from a portfolio perspective and it was a you know build versus partner uh decision and we knew that we could get what we've been working on over the last four five years with NetApp and actually enhance that in a variety of ways which we'll talk about in a little bit and that was really one of the main impetus behind offering this and integrating it into now a Google service that is a Google managed service as opposed to a NetApp managed service. So from a customer experience, Eric mentioned the command line, you can enable us through the console, enable the APIs and begin to use it right away. It's very seamless. It's just like any other storage offering that we have. Um, and it really fits nicely into the portfolio. Uh, because if you look at different application requirements uh for enterprises bringing different workloads to the cloud or even companies that are developing from the beginning in the cloud, block file and object storage is still needed, right? And object storage is not the answer for everything. Uh, and there's a lot of unstructured data. You saw the demo yesterday with Vert.ex AI. Just managing and understanding unstructured data is something that object storage typically doesn't do a good job with. And so we really looked at this as a way to round out that portfolio between cloud storage which is u our object storage uh persistent disk or hyperdisk which is our block storage and we also have uh file store which is an NFS uh managed offering. Uh and we announced uh parallel store which is a parallel file system. But what we didn't have is some of the things that Eric mentioned with SMB, NFS, multi-protocol, and the experience that NAP has particularly with the SMB and Windows workloads and VMware workloads. So that really rounds out our storage portfolio to have the right storage solution for the right workload for the customers. >> Thank you, Sean. And next, let me introduce Guna. GA Marapuri is our leader on GCNV on the NetApp side. Gona started off as a developer at HPE. uh afterwards he went on to lead product management at Marll Q Logic and then also did a storage startup like the rest of us here. We could talk about that afterwards and then was a fellow and architect at Samsung and later Western Digital. And so Gona, you've been part of building the demo uh foruh presentation yesterday where you're showing Vert.ex AI and I realize we're doing a lot from enterprise storage, cloudnative storage and genai is something yet different. What are some of the things that you were hoping the audience or people were taking away from the demo yesterday? What are the challenges? What are the benefits? >> Yeah, sure. Thank you, Eric. Good to be uh on the stage with Sean and Eric. Um before I answer Eric's question, like you know, it's been a fantastic partnership. You know, we launched the product end of August. So, you know, this is um you know, we look forward to all your adoption. So, going to the demo, uh it's kind of bringing the future now. uh you know vitics AI is happening uh we have a lot of data the if you happen to be in this morning's keynote with her you actually saw like there's like a massive I think you talked about 20,000 customers out there using in VMware deployments and there's like a lot of data out there so the idea is like you know how best we can use the power of a tech AI you know I think Nvidia char Charlie I think if I get the name correct you know he's talking about there's so much power you know can I bring the AI now to your user know the AIunused is kind of you know left over. So you all of you have data. So the idea is to show very we believe we just scratched the surface of what is possible you know it's very seamless you can use it. We just want to get to you know feel you the power that we don't need to be intimidated by AI especially with Vitex AI a fully managed service combined with fully managed GCNV you know it's up to you for you to use. >> Great. And so maybe I'llgo into a little bit of a the Q&A portion of this session, right? And what I'd start with is I want to ask thepanelists some of the questions, but one thing is uh if we look at the stat from ESG, it's that 97% of customers are using of their survey respondents are using is today, right? So at least from that representation if that's true then we can extrapolate and say most of the world or most of the enterprises they're using some form of is and it's a question of how much. So maybe to get a sense from the room how many of you are running your storage or running portions of your enterprise in public cloud? A couple hands and how many are looking to increase over the next year or interested in increasing storage adoption in the cloud? A few more hands. Okay, great. So let me start with this group and say you know if 97% arerunning in the cloud right uh obviously there's been an incredible growth over the last 10 years I'll start with Sean and ask Sean right uh is there reasons that people are adopting cloud uh have those reasons changed over the last few years as cloud is becoming more and more ubiquitous>> yeah I think um I think a couple things are accelerating this one is um you know people are often times is one of the most precious resources that organizations have. And so by leveraging the cloud, you have the opportunity to let other people do the work so you can focus on other areas of your business. Um that's often times led to faster innovation. Um you know Thomas uh Curry and yesterday mentioned a few um companies that are using AI to do things differently. You know if you take a look at the way business models have worked in the past you know companies like Priceel line are doing things very differently. Um, you know, uh, Home Depot is another example of a customer that's doing things differently because of the cloud. One of the other benefits of moving to the cloud is not only the managed services that I discussed, but also the opportunity to increase and decrease resources as needed. So, the elasticity of compute storage as needed really makes a difference for different workloads that need to be run uh, ina variety of ways, right? And even something as I don't want to say as simple but as traditional as VMware for the last few decades um you know that is something that it can be run in Google cloud as a managed service you have the opportunity to increase and decrease storage compute based upon applications needing to change over time and so I think the elasticity is one of the big reasons that companies are moving to the cloud and even for some of the more traditional uh file serving workloads right we have customers that are just doing Windows file serving that have been using NetApp on premises but there'sa sometimes there's a better way to do things that you've done in the past. Uh, and that's actually one of the ways that moving file services into Google Cloud. Uh, and actually running that and having that version of Truth, if you will, for those files with wide area file locking, be able to have that in a regional, but then actually be able to replicate that to a second region for availability. Still provide the users the local seamless experience that they have for accessing the files. But now the protection, the management can be rethought where you're not managing necessarily data center facility space. you're not managing and procuring hardware on a three or five year life cycle. Again, that elasticity of what you need and what your users needs change over time, you could adapt to>> andfor good uh what about for you? Are you seeing the reasons people are adopting cloud changing or evolving over the last few years? >> I think across our um customer base yesterday Sean and I were discussing like you know there's like a massive healthcare company uh from Midwest. they actually migrated u to GCP uh primarily because of elasticity. The other thing is you know they actually have a new breed of applications coming enabling them in the cloud is much easier and faster. So I also see some of our customers using for time tomarket reasons.>> I I'll add I'll say from my point of view there's been a couple things. There's been large data center shutdowns where customers at times will move to cloud and often times that's a CIO top down decision. So that's a case where we see a lot of lift and shift workloads happening but for the reasons that Sean and Guna added the idea of how to innovate those are also happening. So I think as we look at our portfolio, one thing I'm seeing as customers are starting to say, how can I get more productivity and also focus on what I'm good at, right? So they're building applications and they're in some ways transitioning. That also creates some audience differences and so maybe a continuation is, you know, I'mthinking that the workloads are changing uh and the workloads might be I'm going to run some SAP here. I'm going to run some Windows VMware elsewhere. On top of that, they're trying to do new and new workloads with things like containers, but either from a migration or adoption, uh more on the workload side, are you seeing differences, Sean, in how customers are adopting these workloads in cloud? >> Yeah, you know, [clears throat] excuse me, the regional availability is one of the big things for and benefits for cloud, right? So, Eric at the beginning mentioned that we've had NetApp cloud volume service and CBO cloud volumes ontap available in 28 regions. We have uh Google has a presence in 38 regions worldwide, right? So this new service that we have with NetApp volumes, we're bringing that out to all the Google regions uh in the first quarter of next year. So this is again one of the other things that we can do based upon the extending the partnership is extend that reach of where customers can actually deploy leverage uh net volumes. But this also points to the other availability of compute resources, right? Whether that be GKE or compute engine or in some cases GCVE the Google cloud VMware engine um getting the data close to the customers depending upon your business is an important aspect and so if you can actually distribute compute resources distribute data across multiple regions where you're not having to have your own data center facilities you can leverage Google's facilities is a way to deliver sometimes a better experience um and typically like content delivery is not necessarily something that is a file service it's more of a cloudservice object storage service but that's one of the reasons why customers will often times choose Google cloud multi-reion object storage or buckets so you can get the data closer as well as if you're doing compute analytics whether it be you know bigquery for traditional analytics or the AI that Thomas and George spoke about yesterday getting the data as close to possible as to the compute finding GPU resources thatare available since you typically can't buy them or afford them yourself those types of things often times lend themselves to designing applications differently than if you were thinking about an on- premises one or two data center for redundancy deployment model. And then one thing as I as we think about how uh there's different compute resources how to access it we've also had a different pattern with being hardware based in our Google cloud offering as well as softwarebased and so maybe for you GA are there things that you think the softwaredefined storage that we have in Google cloud and we're investing in is that going to change uh how we can deliver or reach customers? >> Yeah absolutely. Um uh so if you guys know like we have a hardware offering um that we offer extreme latencies like you know we talk about like you know for like someregions you can actually go out like sub millisecond and most regions will be like 1.4 4 milliseconds. But then where the innovation is happening as he's talking about is like you know Google is expanding to pretty much the whole uh the world like right now 38 as he's saying like uhwhat we observe across our customers you know I'm not going to tell the names but like you know we see it in India we see in APAC not just not only North America there's this customers who want to get to regions local to their users. So whenever they're actually approaching theproblem like that, you know, they want the same enterprise data management that ONAP offers like you know for their lift and shift of applications this could be traditional VM applications using either SMB3 or NFS but they want the application closer to where their users are and that's where we expect like know the software to do give us a massive advantage. It's the same operating environment built oneasily accessible to the customers where they are where they want to deploy applications and very soon like you know theother aspect is Google as an IAS provider keep on improving their hardware infrastructure their compute and uh persistent storage infrastructure. So it gives us a huge opportunity for us to build on those innovations and pass to our customers. >> Yeah and I I'll continue I'll say that the software defined it can be in to Sean's example, not just 26, but all the regions, right? There's places where hardware isa great fit, but there's places where software can let us launch faster. And some of it is also we're seeing a change in our audience. I think that there's a central ro role that storage administrators play, especially in NetApp's traditional use cases. And I think in the cloud that's a place where I think we can innovate because there are cases where customers are migrating to the cloud and they're using their application teams but given this group and this audience I think there's a role that we've yet to fully address which is how does the storage teams participate and drive that and I think with software with the idea of software define being closer in regions that will hopefully give customers more access and more options. It also gives us the ability to address it with more innovation in terms of control plane ability to import. So one thing that I know that the netapp install base cares a lot about is something snap mirror which most of you or some of you have heard right. So uh going to um you're using snap mirror today. How are you using snapmir and how do you plan to use snap mirror? Yeah, there are two aspects. Um you know the key aspect of uh in general like know on tap is resiliency and you know how talked about protect built on snap mirror most of you know today like we're already using snap mirror to provide disaster recovery or asynchronous replication if you guys know snap mirror async between two regions you know uh one of the design patterns of the cloud is there's no single point of failure right and if you have traditional data centers usually have two data centers if you kind of move to a region like you know you could be US west one you know you want to go to US west two or maybe east right um and similarly like you know any of the regions in EMI or APAC so we use snap mirror um to give you the flexibility at the volume level the in contrast with what you can do with a typical ontap SVM you know you can actually do at the volume level like you know that way you can say oh these are critical applications I'm going to provide DR or synap mirror as sync between the regions for some volumes they could be test and dev or could be like non-critical you don't need to replicate you know whatever typical things that you do in the cloud the other thing um we are building on top of snap mirror netap some of you already may know so I'm going to use some of the onap vernacular because I saw you guys are familiar with onap on also has something called snap mirror to cloud so where you actually take snap mirror and put return the waffle format and put in an object storage. So the advantage is when you restore, you know, it can actually restore much faster because the onap format we actually use it for our backup. You know, we just started the preview and uh you know, we're going to build on top of the capability. >> Yeah. And I think what good is starting to say is snap mirror the cloud will allow us to import and move data more quickly. And then on the snapmir side, one thing that customers have brought up to me is that when they have a disaster recovery, they might use something else like a HANA system replication, but even with HANA system replication, they can't move to a different checkpoint. So they need all the systems to coordinate. So with Snap Mirror, you can go restore to not just the most recent, but a prior instance or a prior point in time. So there's a lot of things that the traditional Statmir has provided that we think in cloud will be even more useful. So I thought that was kind of cool. And then on the cross aaz pattern I know Sean you've been talking about cross and the resources uh you know how are you asking customers not just from a storage but how is cross impacting your deployments and going to how's that impacting yoursoftware defined strategy? Yeah. So, you know, we talked a little bit about the hardware and the software models. Uh, hopefully nobody is actually aware of that, right? Because as a service, you should only care about what you're consuming. And that's a different model than thinking about on premises hardware, you know, the cash to spindle ratio and what the cash hit ratios are and doing all the day-to-day things that we've done for many years. That all goes away. And so, if you think about the availability zone question, Eric, it's really about a service level that you're consuming, right? And so we have currently we have a 39 SLA for storage services but we also have a 3 and a half 9 SLA storage service depending upon what region you're in right and so behind the scenes you know the different deployment models of the hardware and software behind that but from what you as a customer see is an experience about consuming a storage service at a dollar price point to performance and an availability uh and that availability is going to increase over time uh where you're going to be able to have like a Fortnite SLA where You actually have regionally available storage that can actually sustain a zonal outage within a region. So in every region there are three zones and you can choose the protection level you want on a perv volume basis. So we're going to talk about snap mirror cloning replication can be to another region. asynchronous. We have two asynchronous file systems available backup to behind the scenes a cloud storage bucket more permanent less costly but longer recovery time than an async replication with cross region replication. So you can think about things more as a service that you want to consume on a per application per volume basis and you can mix and match your priorities based upon your application needs. So the availability zone question is one of what do you want to protect against file system corruption a zonal outage within a regional outage or something else? Yeah, if I add one thing to what Sean said like you as users, you know, we kind of giving you a peak of what is behind the store, but you actually consume service, right? You know, so and at the volume granularity that is the best part of uhNetApp volumes. Um going back to Eric's question on like you know uh again you as a user don't need to worry about you know can I leverage the latest innovations whether that's Net Apps and Google's kind of joint response built-in partnership to deliveryou thefully tuned performance optimizations and latency that we deliver with our AFF systems and also the power of onap in every GCP region. So the Eric talked about snapware as a key capability because uh most of the conversation start with how can I protect there's other aspect you know we haven't talked yet is you also have data centers today on prem you know how can you actually use cloud as a DR target or you know how can you actually migrate the workloads and again you know snap error technologies as a data more play a very critical role >> so maybe uh before I ask one last question right uh I'lljust try to summarize a few things that I want to make sure everyone sees is in 2019 we launched cloud volume service and Google since that time it's been a great partnership and that's included both the awards as well as things that have been important to customers right meeting them where they are helping them with their migration helping themwith their new applications and then this year we launched Google cloud netapp volumes in August and that's only the start there's a lot that's coming and I think the key thing is that the partnership and the innovation as it's reborn in this new first party context. There's innovation happening andSean's right. We don't need to worry about whether it's hardware or software. We're trying to meet you where you are. So whether it's cross aaz cross region, snap mirror protection, the ability to import with the waffle format with snap mirror as well as things that we haven't talked so much about customer manage encryption keys. All of those are trying to help you take advantage of your enterprise workloads or your enterprise uh ability to run in Google with the ONTAP technology. So on top of that, the innovation that's happening all around from aspects that deliver the service, those will only increase. So something that I'd like to just make sure people should see is that the partnership, the growth, it's just started, right? Even these features, they're only the beginning. So maybe with that I'll turn to Sean and then going to ask like forthe audience what's one takeaway you'd want them to have about either GCNV or the partnership. Um I think one thing is that you know as storage technologists we oftenimes think about storage and how we can do things better but in ultimately it's about what you're delivering to theyour business right is it an application and what's supporting that application and this is one of the things that when you're running resources in the cloud you often times have to think about things more holistically as opposed to um just a storage admin that's responsible for delivering a storage service but really what's sitting on top of that and when you look at cloud deployments and even just something I want to say a simple but troubleshooting a problem for performance that an application is seeing. One of the benefits and one of the integration points that we now have with NetApp volumes is integration into our cloud audit, logging, monitoring. So from an administration perspective, you can see in the cloud console all of your monitoring from all of your resources, network, compute, storage in one place, right? And that's different than compared to an on- premises. The storage team has the storage dashboard. The network team has a network dashboard. Compute database, they have their own. And then you have to triangulate and try and figure out what's going on. And so if you think about integration and things that we're going to be continuing to deliver, it's not only the some of the initial audit logging anduh monitoring I talked about, but other Google services like identity access management, right? what storage admins have what permissions to do what and what compute resources can they do based upon Google's identity access management services that we now offer with part of the netapp volumes. So a lot of those other things and maybe even something like billing right that's often times an important thing for customers that you get one bill for exactly what you need and it's uh exactly what you're paying for and that le elasticity allows you to increase or decrease your bill based upon uh changing billing changing business needs. Yeah, onething I want to say is um key takeaway is when you're thinking of cloud adoption especially as an ontop customer there are no tradeoffs you're not using everything what Sean said you know the flexibility elasticity versus any functionality there like know we are in a run to get like everything that you will have on prem will actually show up on GCNV onour net app volumes we already have the critical functionality we have some pretty exciting road map features So any and we would love to partner with you if you have any new applications that you're using, you have some cool features or very specific features of hunap you put to use, you know, please let us know. We'll work with you to uh get them to NetApp volumes uh you know when are you ready to move. >> Sowith that uh I'd like to just ask the audience if they have any questions for us or our colleagues please. >> Thank you. So I'm going to extend this question. Thefirst thing is like uh on tap supports fabric pool. You may be using storage grid on prem you know for the tiering reasons. The question is about what is the plan for fabric pool. So as weg you know we talked about the data protection. The next thing is the one thing that's top of all of our mind is cost optimization for the cloud right you know um so we are it's actually in the works. Um I don't want to give any particular date but you know it'll come very soon. I was one other thing one of the advantages of fabric pool is that it allows you to optimize cost is going to set because in the background we actually will move that to object storage or Google's cloud storage right but you as a consumer of the service don't have to manage a bucket you just decide what percentage of the volume you want to tier based upon a policy of last access of 7 days or 30 days or 90 days and you get totweak those knobs about what's going to move data where however Everything is still accessible through the file system, right? So, Waffle still maintaining all that file access. It's not directly accessible by object storage, but it allows you to bridge that pricing gap between the service level that you want to have for most of the file system access and then something less than that for that amount of storage that's tiered off. And based on the application, you know, you can slide that dial from 20 to 80% if you want. >> Any uh let me make room for one more question. Any other questions for this group? Okay. So, Iwould wrap up by saying I think this has been a great partnership. Thank you, Sean, and to the Google team. I think it's just the start. Please uh I'll show some other talks we're giving, but feel free to stick around and ask our colleagues questions. Thank you all. Thanks for taking time. Thank you. [applause]
With Google Cloud NetApp Volumes, customers can easily migrate apps to Google Cloud without refactoring code or redesigning processes. This fully managed file storage service is based on leading NetApp ONTAP storage-management technology.